9. Machine Learning at the Edge
During the last decade, machine learning required massive servers, complex models, teams of extremely rare experts, and excessive amounts of time and resources. Now, machine learning can fit inside a motion sensor, and we can expect a lot more intelligence at the edge. It will never replace what we can achieve with cloud computing, but it will rapidly complement it. By implementing decision-making systems inside sensors, engineers can optimize resources, save a lot of energy, and time.
- Nomination of the LSM6DSOX, First Inertial Sensor with Machine Learning
- STM32Cube.AI: Convert Neural Networks into Optimized Code for STM32
- SensorTile in a Box: More Powerful Sensors and Three User Modes